An Adaptive Headlight Dimming Systems for Enhanced Night Driving Safety with Machine Learning on ARM Boards

The very fast development of road transport, though very significant to economic and social growth, has also caused a dreadfully huge increase in road accidents particularly night driving. One of the reasons for such accidents is the Troxler effect, which is a visual impairment caused by the glare of oncoming high-beam headlamps that lowers drivers' visibility and response time. This work proposes an embedded system and ML-based intelligent vehicle headlight solution for nighttime hazard reduction. Using real-time sensor feedback and rule-based control on low-power devices such as STM32 microcontrollers, we propose an optimized glare reduction headlight dimming system based on adaptive responses to dynamic traffic conditions. The strategy focus on energy efficiency, simplicity, and cost-effectiveness while offering a practical alternative to computationally intensive complex machine learning algorithms, making it suitable for deployment on resource constrained systems like embedded and IoT-based automotive platforms. Through the integration of embedded and IoT architectures, the research seeks to improve significantly road safety to allow for future smarter, safer, and more responsive automotive systems.